Method and device for extracting specified image subjects

Abstract

Method and device for extracting a specified image subject including an extracting condition of an extracting algorithm of a subsequent stage so as to be adapted to an extraction result by the extracting algorithm of a precedent stage, when specified image subject extracting algorithms are successively implemented. At the time of performing the specified image subject extracting algorithms in each stage by parallel processing, the method and device manage extraction states of each step and qualify extraction processing conditions in the subsequent stage in accordance with the extraction states in the precedent stages. The method and device perform a vote in an N-dimensional space of image characteristic quantity for each extraction area by the specified image subject extracting algorithm and then perform weighting of degree of certainty as the specified image subject based on an aggregation value of the vote within a section area for aggregation in the N-dimensional space.

1. A method of extracting a specified image subject which implements a plurality of specified image subject extracting algorithms in each stage of a plurality of stages by means of parallel processing, comprising the steps of:
 managing respective extracting states of said plurality of specified image subject extracting algorithms in said each stage; 
 qualifying respective extraction processing conditions of said plurality of specified image subject extracting algorithms in a subsequent stage according to the respective extracting states in a precedent stage; and 
 implementing said plurality of specified image subject extracting algorithms of said subsequent stage under the thus qualified respective extraction processing conditions by means of parallel processing, 
 wherein said precedent stage comprises extracting a shape of regular geometric form of the specified image subject. 
 
     
     
       2. The method of extracting the specified image subject according to  claim 1 , wherein said respective extraction processing conditions are areas to be subjected to extraction processing when implementing said plurality of specified image subject extracting algorithms of said subsequent stage. 
     
     
       3. The method of extracting the specified image subject according to  claim 1 , wherein said respective extraction processing conditions are types of extracting algorithms to be implemented at said subsequent stage. 
     
     
       4. The method of extracting the specified image subject according to  claim 1 , wherein said respective extraction processing conditions are control parameters inside extracting algorithms to be implemented in said subsequent stage. 
     
     
       5. The method of extracting the specified image subject according to  claim 1 , wherein said plurality of specified image subject extracting algorithms to be implemented by means of parallel processing in said each stage are of same combination in said plurality of stages. 
     
     
       6. The method of extracting the specified image subject according to  claim 1 , wherein said plurality of specified image subject extracting algorithms to be implemented by means of parallel processing in said each stage are of different combination in said plurality of stages. 
     
     
       7. The method of extracting a specified image subject according to  claim 1 , wherein said extraction processing conditions comprise electronic flash or backlight information. 
     
     
       8. The method of extracting a specified image subject according to  claim 1 , wherein the plurality of specified image subject extracting algorithms in each stage of the plurality of stages are implemented at a same time. 
     
     
       9. The method of extracting a specified image subject according to  claim 1 , wherein said image subject extraction parallel processing units comprises:
 skin color extraction, face contour extraction, hair-on-head extraction, eye/nose/mouth/eyebrow extraction, body extraction, and non-background area extraction. 
 
     
     
       10. The method of extracting a specified image subject according to  claim 1 , wherein said algorithms comprise different degrees of resolution. 
     
     
       11. A method of extracting a specified image subject according to  claim 1 , wherein said regular geometric form is circular or elliptical. 
     
     
       12. A method of extracting a specified image subject according to  claim 1 , wherein said qualifying respective extraction processing conditions of said plurality of specified image subject extracting algorithms in the subsequent stage is performed after the respective extracting states in the precedent stage. 
     
     
       13. A device for extracting a specified image subject, comprising:
 a plurality of stages of image subject extraction parallel processing units, each image subject extraction parallel processing unit for implementing a plurality of specified image subject extracting algorithms in each stage of said plurality of stages by means of parallel processing; and 
 a control unit for managing respective image subject extraction states of said plurality of specified image subject extracting algorithms in each stage by said each image subject extraction parallel processing unit of said image subject extraction parallel processing units and qualifying respective extraction processing conditions of said plurality of specified image subject extracting algorithms in a subsequent stage of an image subject extraction parallel processing unit according to the respective image subject extraction states by the precedent stage of the image extraction parallel processing unit, 
 wherein said precedent stage comprises extracting a shape of regular geometric form of the specified image subject. 
 
     
     
       14. The device for extracting the specified image subject according to  claim 13 , wherein said control unit qualifies areas to be subjected to extraction processing in said subsequent stage of said image subject extraction parallel processing unit as said respective extraction processing conditions. 
     
     
       15. The device for extracting the specified image subject according to  claim 13 , wherein said control unit qualifies types of extracting algorithms to be implemented in said subsequent stage of said image subject extraction parallel processing unit as said respective extraction processing conditions. 
     
     
       16. The device for extracting the specified image subject according to  claim 13 , wherein said control unit qualifies control parameters inside extracting algorithms to be implemented by said subsequent stage of said image subject extraction parallel processing unit as said respective extraction processing conditions. 
     
     
       17. The device for extracting the specified image subject according to  claim 13 , wherein said image subject extraction parallel processing unit implements said plurality of specified image subject extracting algorithms with same combination in each stage of said plurality of stages by means of parallel processing. 
     
     
       18. The device for extracting the specified image subject according to  claim 13 , wherein said image subject extraction parallel processing unit implements said plurality of specified image subject extracting algorithms with different combination in each stage of said plurality of stages by means of parallel processing. 
     
     
       19. The device for extracting a specified image subject according to  claim 13 , wherein said extraction processing conditions comprise electronic flash or backlight information. 
     
     
       20. The device for extracting a specified image subject according to  claim 13 , wherein the plurality of specified image subject extracting algorithms in each stage of the plurality of stages are implemented at a same time. 
     
     
       21. The device for extracting a specified image subject according to  claim 13 , wherein said image subject extraction parallel processing units comprises:
 skin color extraction, face contour extraction, hair-on-head extraction, eye/nose/mouth/eyebrow extraction, body extraction, and non-background area extraction. 
 
     
     
       22. The device for extracting a specified image subject according to  claim 13 , wherein said algorithms comprise different degrees of resolution. 
     
     
       23. A device for extracting a specified image subject according to  claim 13 , wherein said regular geometric form is circular or elliptical. 
     
     
       24. A device for extracting a specified image subject, according to  claim 13 , wherein the control unit for managing respective image subject extraction states of said plurality of specified image subject extracting algorithms in each stage by said each image subject extraction parallel processing unit of said image subject extraction parallel processing units and qualifying respective extraction processing conditions of said plurality of specified image subject extracting algorithms in the subsequent stage of the image subject extraction parallel processing unit is performed after the extraction of the respective image subject extraction states by the precedent stage of the image extraction parallel processing unit. 
     
     
       25. The device for extracting the specified image subject according to  claim 14 , wherein said control unit checks a degree of a overlap of each of candidate areas of said specified image subject which is extracted by a plurality of said specified image subject extracting algorithm, qualifies a candidate area in which the degree of the overlap exceeds a predetermined value as being said areas, and controls said image subject extraction parallel processing units to extract a undetected related information. 
     
     
       26. The device for extracting the specified image subject according to  claim 15 , wherein said control unit checks composing elements of a specified image subject extracted by a plurality of the specified image subject extracting algorithms, and qualifies the types of extracting algorithms to be implemented in said subsequent stage to extract undetected other composing elements.